Perhaps we are gradually entering the era of self-driving cars when a series of “boggarts” in the technology world such as Apple, Google and especially Tesla are holding technological development projects. driverless car technology is extremely ambitious. However, despite the “tons of money” investments and tireless efforts of the manufacturers, fully automatic cars have not really been widely used.
In fact, there are still many fundamental challenges that manufacturers must solve before thinking about creating a truly safe fully automatic car that can pass complex tests. and especially win the trust of users when dealing with real-life situations. Here are 5 reasons why self-driving cars are still not widely used in the world, despite being invested and promoted a lot in the past time.
* Sensor .
Self-driving cars use a range of sensors from basic to advanced to observe and perceive their surroundings in real time, helping them detect objects such as pedestrians, other vehicles and road signs. For example, image sensors (cameras) help the car see objects. Lidar sensors use lasers to measure the distance between objects and vehicles. Radar sensors are responsible for detecting objects, tracking their speed and direction.
All of these sensors collect data and send it back to the car’s control system (usually an AI computer). Here, the data will be carefully analyzed to help the car make the most accurate decision, such as where to steer or when to brake, brake force … A fully automatic car will need. a sensor system that operates without error in all conditions and environments without human intervention.
However, practical testing has shown that factors such as severe weather, heavy traffic, complicated road signs, etc. can all negatively affect the accuracy of the sensor. The radar used by Tesla cars is less susceptible to adverse weather conditions, but it remains a challenge to ensure that the sensor can detect all objects with the degree of certainty needed to ensure that all objects are detected. absolute safety for the occupants of the vehicle.
And in crowded cities, there is a complicated traffic situation like Cairo or Hanoi.
* Artificial intelligence .
With that said, most autonomous vehicles will use artificial intelligence and machine learning to process the data obtained from the sensor system and make decisions specific to each situation. This can be compared to the brain of the car.
AI algorithms are trained to identify the objects detected by the sensor, then classify them correctly. Next, the computer uses this information to decide if the car needs to take action, such as braking or turning, to avoid an object.
In the future, machines will be able to do this detection and classification job more efficiently than humans can themselves. But for now, there’s no guarantee that the machine learning algorithms used in cars are completely secure. There is a need for a standardization of how self-driving machine learning systems should be trained, tested, or validated.
* Stabilization .
When a self-driving car rolls on the road, it’s constantly learning, taking new roads, detecting objects it hasn’t encountered during training, and forcing periodic software updates.
How can we ensure that the system continues to be as secure as the previous version? Or whether there are errors or vulnerabilities that appear after a software update affects the vehicle’s ability to operate. Every mistake no matter how small in this situation can lead to catastrophic accidents.
* Standards .
There is still no unified system of international standards and regulations for autonomous vehicle technology. For this new type of vehicle, new regulations are required for specific functions, such as for automatic lane keeping systems. Since then, car manufacturers are forced to comply before the product is licensed.
This is an issue not only related to safety, but also to a variety of other aspects such as environmental, economic and social.
* Social acceptance .
There have been many accidents involving self-driving cars in general and of Tesla in particular.
Everyone needs to be involved in decisions about the introduction and adoption of self-driving vehicles. simple andJoining traffic with artificial intelligence systems is unprecedented and people have the right to question the safety of the community.
The first three challenges must be addressed to help overcome the latter two. Of course, the race in the field of self-driving cars will not cool down. But without a sit-in between manufacturers, consumers, testing agencies and regulators, self-driving cars will probably remain on test roads for years to come.